1. Field of the Invention
The present invention relates to a Raman amplifier system, apparatus and method for amplifying optical signals in various optical media, and more particularly to a Raman amplifier system, apparatus and method that employs a control unit for achieving an arbitrary Raman amplification profile in a single Raman amplifier as well as a plurality of cascaded Raman amplifiers in a communications network.
2. Discussion of the Background
Optical fiber communication systems transmit optical signals over considerable distances. However, the transmitted optical signals are attenuated because of absorption and scattering, ultimately resulting in signal degradation. To keep the level of signal strength above background noise by a predetermined amount, the optical signals must be periodically amplified. Typically, optical signals are amplified using electronic repeaters, which convert the optical signals into electric signals, amplify the electrical signals, and then convert the amplified electrical signals back into optical signals for further transmission along an optical fiber.
Since 1996, network traffic growth has continued at a torrid pace, with Internet traffic doubling every 100 days and overall network traffic growth approaching 100% annual growth. In an effort to keep pace with bandwidth demands nearly doubling annually, wavelength division multiplexing (WDM) system designers are moving to more cost effective solutions such as optical amplifiers.
Generally, there are two general types of optical amplifiers for amplifying signals within optical fiber communication systems. The first type is a rare earth doped fiber amplifier, such as an Erbium Doped Optical Fiber Amplifier (EDFA) using Er (erbium) doped fibers as an amplification medium. The second type of optical amplifier is a Raman amplifier.
EDFA is currently the most widely used optical amplifier for WDM systems and is effective and reliable in optically amplifying WDM signals. However, an amplification bandwidth of EDFA has a limited range of about 1530 nm to 1610 nm. Further, as shown in FIG. 1, EDFA produces a wavelength dependent gain profile with a peak gain between 30-36 dB in the 1525 nm to 1540 nm range and a more flat gain plateau at about 30 dB in the 1540 nm to 1560 nm range.
Accordingly, when EDFA is used to amplify WDM signals, which are spectrally distributed over the amplification bandwidth, a non-uniform amount of gain is applied to the separate WDM channels, depending on the wavelength of the channels. To offset this effect, a gain flattening filter is used to obtain a uniform or flat gain profile (a gain deviation of less than 1 dB) across the entire communication band. In particular, a loss profile of the gain flattening filter is designed to have an inverse shape similar to the gain profile. However, the filter is limited to a particular gain profile and is not dynamically adjustable to compensate for changes in a magnitude of the gain of the EDFA. Accordingly, a flat gain profile cannot be maintained when the gain of the EDFA is changed. In addition, the gain flattening filter decreases the total amount of power launched into an optical fiber.
Further, WDM systems using EDFAs are plagued by noise problems. In more detail, there are two types of noise associated with fiber optic transmission: span noise and amplifier noise. In a system amplified exclusively by EDFAs, a signal leaves a first amplification site at a high end of the system's dynamic range. However, over the next approximately 75 Km, the signal attenuates linearly and reaches the next EDFA at a level that is much closer to the background noise floor (i.e., lower signal-to-noise ratio) than when originally generated. Thus, when the EDFA amplifies the signal, the EDFA amplifies both the signal itself as well as the background, such that the signal-to-noise ratio is not further degraded by background noise when transmitted further down the optical fiber. The amplifier itself, however, introduces some amplifier noise into the the signal and thus the signal to noise (background plus amplifier noise) ratio decreases after each EDFA amplification stage.
After about 400-600 km, the signal has to be regenerated (i.e., “cleaned up”). To accomplish this, regenerators are strategically situated throughout the network. The regenerators convert the optical signal back to its electrical equivalent so the data may be detected, as is done in a receiver. Then, the data signal is converted into an optical signal and retransmitted on the network. However, requiring regenerators to be positioned every 400 to 600 km is a costly process, which accounts for up to 50% of the total cost of a network. Regenerators are also very expensive which further increases the total cost of the network.
Raman amplifiers use a phenomenon known as Stimulated Raman Scattering (SRS) of light within an optical fiber to achieve a gain in a particular wavelength band. SRS produces a peak gain at a frequency which is smaller than a frequency of the light pumped into the optical fiber by about 13 THz (or conversely produces a peak gain at a wavelength which is longer than a wavelength of the light pumped into the optical fiber by about 100 nm). For example, FIG. 2 illustrates a gain profile resulting from the use of a pumping device including a single semiconductor laser with a central wavelength of 1450 nm. In this case, the peak of the gain profile is at approximately 1550 nm (i.e., shifted 100 nm from the 1450 nm central wavelength of the semiconductor laser) and the profile has a bandwidth of about 20 nm within a gain deviation of about 1 dB.
However, Raman amplification has primarily been investigated for applications in wavelength bands that can not be amplified by EDFA, because the Raman amplifier requires a greater pumping power to obtain the same gain as that of the EDFA. Thus, traditionally Raman amplifiers have not been used to amplify WDM signals, but this may be changing more recently.
In conventional optical communication systems, optical amplifiers for WDM signals are a basic system component that, in combination with the other system components, define the system's communication performance. When establishing new communication systems, or when upgrading existing systems, system operators perform capacity allocation analyses that determine the number of amplifiers that are required to keep the signal above the background span noise when transmitting the signal from one location to the next.
Due to the rapid growth of Internet traffic, system requirements change frequently. When changes occur, the system analysis must be readdressed to ensure that the components that have already been fielded are able to handle the change in system requirements. Since the amplifiers that are fielded have a predefined system performance, with regard to gain shape across the amplification band, it is generally not practical (from a cost and complexity perspective) to provide field upgrades to the amplifiers to alter their gain profiles to optimize system performance at minimal cost. The common solution therefore is to place additional amplifiers at intermediary points between two already fielded amplifiers, or simply replace the fielded amplifiers with more capable amplifiers, albeit at high expense.
Because amplifiers are conventionally considered to be a discrete component of a larger network, when fault conditions occur at a particular amplifier, the repair action is typically taken only on that amplifier, without considering whether the repair action can be avoided by using excess amplification capacity within that amplifier, or at adjacent amplifiers (upstream or downstream of the damaged amplifier). Furthermore, changes in network architecture that may effect bandwidth, for example, or by using other fibers, may change the original design premise on which the original communication system was developed. For example, perhaps a new type of fiber is laid between two existing EDFA or Raman amplifiers, where the bandwidth-attenuation characteristics for that fiber are different than for the one it replaced. In this condition, the gain profile of the amplifiers may not be matched to the fiber, thus giving rise to suboptimum utilization of system resources and/or system performance.